Crowdsourcing Adverse Test Sets to Help Surface AI Blindspots

To keep the competition focused, the participants of the CATS4ML challenge will search for adverse example images only in the target images and for the target labels provided below:

  • Target images: These are the subset of images from Open Images Dataset (OID). Only image IDs from these target images will be considered for this challenge.
  • Target labels: a list of the labels for which participants should find adverse examples (AI blindspots).

Check Participate to learn how to start contributing.